Based on a common understanding of what machine Learning (ML) can offer. a lot of value will be created on ArtificiaI Intelligence (AI) features on balance routers for Intelligent Balance. in both the router and the Incontrol2 Console.
AI or ML, gives the extra steps that are missing today on balance routers. Today we have excellent data capture and visualizations that help us understand WAN Health. but no more than that. the right path is to evaluate and reflect WAN Health over time, As an Intelligent Thermostat can do. and then Identify in REAL TIME what should be the distribution of packet through each WAN link and dynamically weighting each WAN Link.
Required steps to achieve AI:
Capture Data (checked)
Store Data (checked)
Analytic Tools (checked)
React (Missing) required tech - Math, signal processing, etc.
Predict (Missing) - required tech - Machine Learning
As Balance Routers are capturing a lot of data continuously, doing the analysis based on historical data will not be a major issue. but one thing should be taken into consideration. every time you apply a chance on the system you should be able to learn from previous data, new data or both.
The features will be both in Balance Routers and Incontrol console. doing different things. Balance router will be confined to the customer’s network and Links, but Incontrol could do further analysis on multiple systems and WAN Links for multiple locations. As it gets broader access to data on the network.
Example. Incontrol notice on a specific problem with an ISP link on an Area it could lower the link health weighting for all balance routers that use this link, while this problem exists.
Also as more historical data, we collect the better we can target problems and possible solutions could be recommended.
I hope this can help get an Idea of how Peplink and user can benefit from AI and ML on this new bread of technologies. And would like to wake up some interest around this topic.